1. Semantic Interoperability in Healthcare State of the Art in the US (position paper) March 15, 2010 CCIB Convention Centre, Barcelona, Spain
W. Ceusters and B. Smith
1 Ontology Research Group, CoE in Bioinformatics & Life Sciences, SUNY at Buffalo

2. Semantic Interoperability
One definition
Semantic Interoperability
the ability of two or more computer systems to exchange information in such a way that the meaning of that information can be automatically interpreted by the receiving system accurately enough to produce useful results to the end users of both systems.

3. No semantic interoperability without standards
There are many standards that are claimed to facilitate SI;
But there are very few that really do so at an acceptable level …
… if we take ‘acceptable’ seriously.

4. Important notions in the context of ‘acceptable’
Semantic Interoperability
the ability of two or more computer systems to exchange information in such a way that the meaning of that information can be automatically interpreted by the receiving system accurately enough to produce useful results to the end users of both systems.

5. Who sets the standards for these notions ?
Vendors and commercial IT developers ?
tend to promote what their systems are capable to do, blocking anything that requires modification
US health IT systems have very old foundations
Government ?
typically very minimalistic population-based and cost-oriented view
strongly influenced by business lobbying
Clinicians, practitioners ?
usually think in terms of ease of use, thus little or no attention for structured reporting
Scientists ?
tend to think in terms of what is possible tomorrow

6. Current barriers (1)
too large a number of players (clinicians, patients, payers, industry, government …) with competing agendas
insufficient coordination based on a shared set of coherent principles
overestimation of the value of terminologies and concept-based ontologies
inadequacy of current systems, primarily electronic health record systems, to capture data adequately for example in ways that support continuity of care

7. Current barriers (2)
inconsistent and badly documented standards, some of them maintained by consultants and others who benefit from inadequate standards and from poor documentation;
shortage of trained personnel who can span the divide between IT and biological and clinical expertise
too rapid turnover of trained personnel, so that promising systems are abandoned or poorly maintained

8. Current barriers (3)
vague, minimalistic notions of ‘meaningful use’ which have little to do with semantic interoperability
the rush – and available government funding – to install existing EHR systems none of which has the right foundations to support full semantic interoperability.

9. Meaning of information:
Need for progress (1)
Meaning of information:
should pay more attention to what information is about, rather than how information is structured or communicated
stop confusion between terminology and ontology
Useful results:
each piece of data entered somewhere for some purpose should be useful at various levels of granularity
from individual patient care to world population

10. Need for progress (2) End users:
should not just be people but also software agents:
decision support
prediction
prevention

11. Conclusion
Although semantic interoperability is a hot issue, not very much has been accomplished.
We accept that the perfect is the enemy of the good.
It is necessary to remove the barriers that prevent scientific progress to be applied in existing systems